"zpool add -n" report is incorrect

[TangoAlpha]

System information

Type	Version/Name
Distribution Name	Debian
Distribution Version	Buster + Proxmox
Linux Kernel	5.0.21-1-pve
Architecture	amd64
ZFS Version	0.8.1-pve2
SPL Version	0.8.1-pve2

Describe the problem you're observing

Output from "zpool add -n pool log logdev" shows logdev added to main pool and not under logs

Describe how to reproduce the problem

1. Create pool without a log device
2. zpool add -n test log logdevice
3. zpool add test log logdevice
4. zpool status test

Output from 2 (incorrect) does not match output from 4 (correct)

Include any warning/errors/backtraces from the system logs

Example:

root:~# zpool create test scsi-36589cfc00000083a8f69ec6f96c766a6
root:~# zpool add test cache scsi-36589cfc0000007fd9df5a667cfa6b459
root:~# zpool status test
  pool: test
 state: ONLINE
  scan: none requested
config:

	NAME                                      STATE     READ WRITE CKSUM
	test                                      ONLINE       0     0     0
	  scsi-36589cfc00000083a8f69ec6f96c766a6  ONLINE       0     0     0
	cache
	  scsi-36589cfc0000007fd9df5a667cfa6b459  ONLINE       0     0     0

errors: No known data errors
root:~# zpool add -n test log scsi-36589cfc00000079787a155ff1f6d5996
would update 'test' to the following configuration:
	test
	  scsi-36589cfc00000083a8f69ec6f96c766a6
	  scsi-36589cfc00000079787a155ff1f6d5996
	cache
	  scsi-36589cfc0000007fd9df5a667cfa6b459
root:~# zpool add test log scsi-36589cfc00000079787a155ff1f6d5996
root:~# zpool status test
  pool: test
 state: ONLINE
  scan: none requested
config:

	NAME                                      STATE     READ WRITE CKSUM
	test                                      ONLINE       0     0     0
	  scsi-36589cfc00000083a8f69ec6f96c766a6  ONLINE       0     0     0
	logs	
	  scsi-36589cfc00000079787a155ff1f6d5996  ONLINE       0     0     0
	cache
	  scsi-36589cfc0000007fd9df5a667cfa6b459  ONLINE       0     0     0

errors: No known data errors
root:~#

[inkdot7]

Similar issue from #9627:

When adding a special device in dry-run mode (zpool add -n), the special marker is not in the list of the would-be configuration:

# zpool add -n znimbus1 special /dev/vde1
would update 'znimbus1' to the following configuration:
        znimbus1
          vdb
          vdc
          vdd
          vde1

Expected result (which is like what zpool status prints, after actual adding):

# zpool add -n znimbus1 special /dev/vde1
would update 'znimbus1' to the following configuration:
        znimbus1
          vdb
          vdc
          vdd
        special
          vde1

Also, the zpool add -n command did not warn that the pool is missing the allocation classes feature, as the actual command (without -n) does:

# zpool add znimbus1 special /dev/vde1
cannot add to 'znimbus1': pool must be upgraded to add these vdevs

Describe how to reproduce the problem

zpool add -n pool special vdev